After upgrading from 2.8.x (maybe it was 2.9.x) to 3.1.2, all the references inside posts to filenames (usually images) which contain accented chars stopped working.
Before, filenames that are displayed in the filesystem like "EXPRESSÃƒO.jpg" would be correctly called out in the post content HTML as "EXPRESSÃƒO.jpg". WP 3.x decided to convert those references to "EXPRESSÃO.jpg" (which is what the filename looked like when created/uploaded by the blog author), so now all the images that have such chars are broken links.
ONLY references to filenames are mangled. All other accented text works fine. i was able to, with the help of a perl script via SSH, obtain a list of all the files that have special chars.
Tabel with all the mangled chars: http://pastebin.com/MMbjJphU (BTW which table is this?)
wp-config shows (yes i already tried to comment them out)
define('DB_CHARSET', ''); define('DB_COLLATE', '');
DB Collation in effect: latin1_swedish_ci default DB Collation: latin1_swedish_ci Tables Collation: latin1_swedish_ci Columns Collation: latin1_swedish_ci
whats the quickest way to solve this problem? from what i see, i have 2 options:
A) i could rename the files to use the correct spelling (EXPRESSÃƒO.jpg -> EXPRESSÃO.jpg). I am leaning towards this. Maybe someone could help with a php/perl/python script that would rename the files?
B) i could batch update wp-posts in phpmyadmin to start using all those weird chars (EXPRESSÃO.jpg -> EXPRESSÃƒO.jpg).